Closed adrtod closed 7 years ago
Are you sure it is not just due to not having been rebuilt in the fewer than seven days since dplyr 0.7.0 was released?
Yes that's right ;) Version 0.6.0 was out on may but never published on CRAN, hence my confusion. I just have to use the devel then!
@adrtod quick postscript:
re bioconductor: The tidyverse
image includes all suggested as well as required dependencies of tidyverse
and devtools
. Since the latter suggests a bioconductor package we need those repos. There are of course no MRAN snapshots for bioconductor, since it already uses versioned releases based on the R version you're running. Like Dirk says this has nothing to do with dplyr
;-)
re rebuild dates: That's actually a good idea -- as you've noticed devel
version is built nightly, but the other versions are only re-built once a month (they are mostly stable anyway so rebuilding is just a way to pick up security patches to old debian sources). But we can probably add latest
to the nightly builds without issue.
re versions: no need to switch to devel
just to get the latest dplyr
; latest
uses the RStudio CRAN mirror, not an MRAN mirror, so update.packages()
or install.packages()
on latest
will set you right.
re dpylr 0.6.0: correct, in the new versioning scheme odd-numbered releases signify breaking changes, see https://blog.rstudio.org/2017/06/13/dplyr-0-7-0/. (Why the odd-number rule instead of http://semver.org/ I have no idea).
@cboettig things are very clear now, thanks!
For instance the version of dplyr installed is 0.5.0 while the version on cran is 0.7.0. Apparently, that comes from the installation of bioconductor in the Dockerfile.
Why is bioconductor needed anyway?